Corey, On Fri, Jun 20, 2008 at 1:45 AM, Corey J Ashford <[EMAIL PROTECTED]> wrote: > Hi Stephane, > > Working on the POWER6 port to the minimal perfmon2, I ran across a problem > that there is a gap in the syscall numbering, and on POWER arch this is > causing a problem because the check scripts expect that the numbering is > monotonic (no gaps). > The reason for the gap is that I wanted to maintain backward compatibility for existing binary tools. I am surprise nobody on LKML complained about this but I guess they are more interested by the low level stuff at this point.
When I introduced those gaps (for the multiplexing-related syscalls), I wondered why was it that pfm_unload_context() was put last. I must admit that at this point I am tempted to get rid of the gaps and move pfm_unload_context(). Given that this is experimental code anyway, we could hack something in libpfm for now. > The two solutions that come to mind are to introduce a set of > pfm_reserved_1, pfm_reserved_2, etc. system calls, so that the gap is filled > in (which then means I need dummy routines defined somewhere). Or I could > just compress out the gap, and add the other syscall numbers later, possibly > after other newly-introduced, non-perfmon syscalls. > > A third solution requiring changing the scripts to be tolerant of gaps > sounds like an uphill battle, so I'd rather not go in that direction. Don't > > What do you think? Personally, I'm leaning in favor of pfm_reserved_1, etc., > but we may have a difficult time getting that accepted by the LKML folks. > If that works around your problem then that's fine for now. ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
